Fig 2: Hbp1 ablation enhances immune response gene programs and impairs cellular response to stress in the cerulean-induced AP model. (A) Schematic diagram of Visium spatial transcriptomic analysis of pancreata derived from HBP1 WT and CKO mice on days 1 and 2 after cerulein treatments (n = 3, each group). (B) Uniform Manifold Approximation and Projection (UMAP) displayed 8 clusters of integrated samples. (C–E) The expression of marker genes corresponding to pancreatic acinar cells (Gm5771, Zg16, Cel, Reg3d, Ero1lb, Rnase1, Rap1gap, Ctrl, Pnlip, and Prss3) from the top genes in group 2 (C), islet cells (Ins2, Ins1, Iapp, Chga, Sst, and Gcg) from the top genes in group 6 (D), and macrophages (Arg1, Ccl6, Fn1, Pf4, Ctss, Thbs1, Ccl9, Apoe, Stab1, and Plin2) from the top genes in group 3 (E). The X-axis indicates each cluster. The Y-axis indicates log2-transformed normalized UMI counts per spot. (F) Mapping of proinflammatory genes in UMAP. HBP1 WT or CKO pancreata are displayed in UMI. Each dot represents each spot expressing cytokine genes, displayed in log2-normalized UMI counts. (G–I) The top 5 GO terms were identified for upregulated genes in group 2 cells on day 1 (G), genes upregulated across all groups at both time points (H), and genes downregulated across all groups at both time points (I), in HBP1 CKO compared with WT pancreata (STRING DB30; threshold: FDR <.05, signal >0.50, strength >0.5 for G, H and 0.25 for I, see Methods). The signal on the X-axis is defined as a weighted harmonic mean of the observed/expected ratio and −log (FDR). (J–K) The expression of genes involved in the cellular response to oxidative stress (Reg3a, Selenop, Lonp1, Rela, Trp53inp1, Sdc1, Prdx4) (J) and autophagy pathways (K) was significantly downregulated in HBP1 CKO compared with WT pancreata (∗FDR < .05). (L) Top5 GO of DEG upon HBP1 expression in iHBP1 OE PDEC (FDR<.05, GOC DB31). (M) Expression of autophagy genes among DEG increased as HBP1 levels increased (∗, FDR<.05). (N–O) Western blot analysis of autophagy markers p62 and LC3 in iHBP1-OE PDEC cells treated with doxycycline (1 μg/mL, 24 hours) to induce HBP1 overexpression, followed by CQ (12 μM, 24 hours) treatment (N-representative images). Quantification of p62 levels normalized to GAPDH and LC3-II/LC3-I ratio was performed. Band intensities were measured using ImageJ. Statistical analysis was performed using Welch’s t-test (n = 4 for p62, n = 2 for LC3) (O).

Fig 3: Generation of pancreatic Hbp1 CKO mice. Relevant to Figure 4. (A) A schematic diagram of a part of the Hbp1 locus (chr12:31920507-31956193) of Hbp1flox/flox mice. One loxP site is integrated into the intron 1 and another loxP and Frt sites are integrated into the intron 2. After crossing with p48Cre mice, cre recombinase deletes the DNA between the 2 loxP sites in Hbp1flox/flox mice and generates p48Cre; Hbp1flox/WT. To make homozygotic p48Cre; Hbp1flox/flox, the heterozygous mice are repetitively bred with each other. (B) Representative agarose gel images to verify the insertion of loxP and loxP/FRT sites in intron 1 and intron 2, respectively. Using Hbp1 loxP site primers in intron1, Hbp1 WT and floxed alleles give rise to 173 bp and 219 bp, respectively. Using Hbp1loxP FRT site primers in intron 2, Hbp1 WT and floxed alleles produce 184 bp and 272 bp, respectively. (C) Confirmation of missing the part of Hbp1 mRNA encoded by the exon 2 using real-time PCR. Real-time PCR primers are designed to detect parts of the Hbp1 mRNAs encoded by exon1/2, exon6/7, and exon 10-11. Duodenum and pancreatic tissues from control mice with intact Hbp1 alleles are used as negative and positive controls, respectively. The Y-axis shows Hbp1 mRNA levels relative to Gapdh. The pancreata from CKO mice lacking Hbp1 alleles show the complete absence of Hbp1 exon1/2 mRNAs, which is a level relatively similar to negative controls, compared with control mice with intact Hbp1 alleles (∗P < .005, Welch t-test, n = 6 technical replicates). (D–G) Confirmation of the absence of HBP1 protein in HBP1 CKO mice. Representative H&E (D, F), HBP1 IHC (E, G) on pancreatic islet cells from control and CKO mice. (H) Amylase and CK19 staining on pancreata from HBP1 WT control (upper) and HBP1 CKO (lower) mice on day 0, 1, 2, 7 after cerulein treatment (n = 3).

Fig 4: The expression of HBP1 shifts from pancreatic islets to acinar cells in pancreatitis. (A) Schematic diagram of cerulean-induced AP model. (B–E) Representative images of HBP1 immunostaining in the pancreata of one set of mice on day 0 (B, inset b’), 1 (C, inset c’), 2 (D, inset d’), and 7 (E, inset e’). n = 3 each time point, images from the rest of the mice are shown in Figure 2. (F–K) Representative H&E images (F, H, J) and IF images of HBP1 expression (G, I, K) on normal-looking acinar (F, G), injured acinar cells (H, I), and PanINs cells (J, K) from patients with pancreatitis. Scale bars represent 100 μm (top images) or 10 μm (bottom images). (L) Summary of HBP1 expression in human pancreatic tissues (311 samples from 49 patients with pancreatitis and 16 normal subjects). HBP1 expression is significantly higher in injured acinar cells and PanINs compared to stromal cells or normal controls (∗P < .05; ∗∗P < .001; unpaired Welch’s t-test; Supplementary Table 1D). (M–O) ROC curves and AUC values of HBP1 expression for each cell type (M, O) and for the combined cell types (N, O) in pancreatitis patient tissues, which were analyzed against the presence of pancreatitis with cancer.

Fig 5: Representative images of immune staining for cleaved caspase-3 (A–D, arrow; positive signals) and IF for autophagosome marker proteins LC3 (E–H) and p62 (I–L) in pancreata from HBP1 WT and CKO mice on day 1 and day 2 after cerulein treatment (replicates n = 3).Arrows indicate positive puncta signals, while dashed arrows denote diffuse cytoplasmic expression.